The present invention relates to data center infrastructure, and more particularly, this invention relates to exchanging tunnel information over an overlay network.
Network virtualization is an emerging data center and cloud computing trend which aims to virtualize a network as seen by end stations in a way that greatly simplifies network provisioning in multi-tenant environments, as well as traditional environments. One of the more common techniques of achieving network virtualization is to use network overlays, where tunnels are established between servers, edge network switches, and gateways to which end stations connect. The tunnel is actually implemented by encapsulating packets transmitted by a source end station into an overlay header that transports the packet from the source switch to a target switch in user datagram protocol (UDP) transport via an internet protocol (IP)-based network. The overlay header includes an identifier (ID) that uniquely identifies the virtual network. The target switch strips off the overlay header encapsulation, UDP transport header, and IP header, and delivers the original packet to the destination end station. In addition to this tunneling mechanism, the edge switches participate in an address discovery protocol, which may be learning/flooding based, or lookup-based.
Overlay networks like Virtual eXtensible Local Area Network (VXLAN) connect geographically separated Layer-2 (L2) networks using tunnels. These are L2 over Layer-3 (L3) tunnels. L2 packets originated by a virtual machine (VM) in a VXLAN and destined to another VM or group of VMs in same VXLAN in another physical location are carried over L3 tunnels.
All unicast packets to known destinations are carried in unicast L3 packets by the tunnel end points. All broadcast packets, like address resolution protocol (ARP), standard L2 multicast packets, and unicast packets to unknown destinations are encapsulated in multicast L3 packets.
In order to support overlay networks like VXLAN, it is advantageous to support IP multicast and multicast routing protocols, like protocol independent multicast (PIM)-sparse mode (SM).
One conventionally used solution to providing address discovery protocol involves performing ARP functionality at the so-called virtual switches, which are actually software entities that reside within Hypervisors. Virtual switches connect multiple virtual end stations, e.g., devices like VMs, to each other and to the physical network.
However, the use of ARP to discover addresses has its drawbacks. Particularly, if an ARP request is originated from a VM, the ARP request is sent as a multicast packet using the multicast IP address through tunnels to all devices in a particular virtual network. Then, the device which recognizes the address referenced in the ARP packet responds to the ARP request indicating the device's media access control (MAC) address. In this way, the source MAC address of the destination device and the source IP address for the tunnel end points are learned by the device which sent the ARP request, and may be mapped together so that subsequent packets destined for this device may be sent without ARP. However, this involves the unnecessary step of repeating this flooding through the tunnel for each ARP request and for forwarding all packets having an unknown unicast destination address associated therewith. This results in a complex deployment of IP multicasting in IP networks which utilize network overlays.